📂 Category
LLMs & Generative AI Articles - Complete AI News Archive
951 articles in this category • Page 1 of 10
- 1. Understanding JSON Mode, Function Calling, and Structured Output in LLMs
- 2. Audit AI tools: inventory, VPN/zero-trust, continuous fingerprinting
- 3. Building and Deploying Custom hipBLASLt Libraries on AMD Instinct GPUs
- 4. Adobe adds AI Assistant to Photoshop, Premiere, Illustrator, InDesign, Frame.io
- 5. Claude Fable (Mythos) 5 shows limited bug‑finding and refactoring aid
- 6. Helion adopts LFBO with on‑the‑fly Random Forest for autotuning
- 7. NAVI‑Orbital performs first in‑orbit autonomous vision‑language inference
- 8. TurboQuant and OSCAR vie in KV cache compression race at ICLR 2026
- 9. Study probes if language models can hypothesize new math structures
- 10. NVIDIA XR AI Enables Real‑Time Multimodal Agents for AR Glasses
- 11. PrologMCP Launches as Task-Agnostic Open-Source Server for LLM Agents
- 12. Reconfigure OpenClaw on Mac Mini to Deploy a Local LLM Model
- 13. Roadmap to LLM Engineer in 2026: Foundations, Prompting, Fine‑Tuning, Alignment
- 14. Attention output GEMM reduces blended Fprop speedup to 1.47× in NVFP4 training
- 15. Estonian institute benchmarks AI models' vulnerability to Russian propaganda
- 16. Study quantifies AI agent trust formation, breakage, recovery in survival game
- 17. UP‑NRPA Allows Dynamic Customization of Dialogue Strategies Without Offline RL
- 18. Mobile NPU powers on‑device diffusion LLM with Multi‑Block Speculative Decoding
- 19. Orchestra‑o1 Enables Efficient Omnimodal Agent Collaboration
- 20. Vision LLMs Expand PDF Parsing to Charts, Diagrams, and Tables
- 21. Claude Fable 5 beats GPT‑5.5 by 13 points on FrontierMath tier‑4 tests
- 22. German Court Holds Google Liable for False AI-Generated Overviews
- 23. Google's DiffusionGemma: open diffusion model for faster text generation
- 24. Google sues Chinese Outsider Enterprise for Gemini-driven phishing on Telegram
- 25. PersonaDrive conditions VLA agents on human driving demos for simulation
- 26. ToolSense Framework Audits LLM Tool Knowledge Beyond Constrained Decoding
- 27. Gemini Omni adds AI video generation, using compute limits based on complexity and size
- 28. Xiaomi's MiMo Code beats Claude Code on 200+ step tasks, free MiMo Auto to V2.5
- 29. OpenAI hires Sottiaux in 2024, shifts from internal tools to ChatGPT overhaul
- 30. Low Kruskal-Rank Adaptation Shows Matrix Rank Stays r, Kruskal Rank Falls to 1
- 31. Anthropic apologizes for invisible guardrails on Claude Fable, first Mythos model
- 32. AI pre‑mediation matched professional mediators in multi‑issue negotiation test
- 33. AVLLMs Mirror VLM and VideoLLM Sequential Flow in Audio‑Visual Tasks
- 34. vLLM uses custom GPU kernels, TorchInductor and CUTLASS for portable inference
- 35. Claude Fable declines basic biology queries; Opus 4.8 responds
- 36. Run DiffusionGemma on NVIDIA GPUs for high‑throughput text generation
- 37. SynIB Introduces Information Bottleneck to Boost Multimodal Synergy
- 38. Understanding AgentOps: Discipline and the agentops.ai Platform Explained
- 39. Grab, CJ ENM, LiveKit praise Gemini 3.5 Live Translate for quality and accuracy
- 40. Apple's top AI concept mirrors vibe coding, using Shortcuts as a model
- 41. CoCoNuT paradigm expands residual stream for latent‑space, multi‑path reasoning
- 42. OmniMem adds modality-aware memory allocation for audio‑visual LLMs
- 43. PathoSage Introduces Three‑Stage Framework for Patch‑Level Pathology Reasoning
- 44. Apple unveils third‑gen foundation model, AFM 3 Cloud shows 36% boost
- 45. NVFP4 recipe speeds JAX/MaxText training on NVIDIA Blackwell and Rubin
- 46. Weaker LLMs Accidentally Delete Content, Shrinking Documents Over Time
- 47. Four New Specific Techniques to Boost Productivity with Claude Code
- 48. Jensen Huang sees token market segmenting into distinct value tiers
- 49. OpenAI to revamp ChatGPT, shift to business customers, rival Anth
- 50. MLP Networks Fit High-Frequency Functions One Oscillation at a Time
- 51. SafeGene Introduces Reusable Safety-Adapter for Cross-Task Model Families
- 52. FAIR-Calib Introduces Two-Stage PTQ Framework for Diffusion LLM Quantization
- 53. Elmes* Automates Fine-Grained Rubric Building for LLMs in Niche Education
- 54. Lean4Agent launches FormalAgentLib to model and verify workflow consistency
- 55. Study Finds No One-Size-Fits-All Strategy for Multi-Agent Communication
- 56. xAI used Anthropic’s Claude via personal accounts after access revoked for months
- 57. Study examines temporal preference concepts in large language models
- 58. Errorquake-10k Benchmark Scores 10,000 LLM Responses on 0-4 Severity Scale
- 59. Three SpaCy Tricks Speed Up Production-Grade Text Processing
- 60. Zhipu AI employs Muon Optimizer and Muon Split in GLM-4.5 and GLM-5 pretraining
- 61. Anthropic says Claude writes >90% of its code; AI pause button urged
- 62. Choosing AI Models: Prioritize Real‑World Needs Over Benchmark Rankings
- 63. ELI releases LLM benchmark showing top models resist Russian propaganda
- 64. AI trust certification trial in Fintech, Banking, Insurance, Health, US, Vietnam
- 65. SMAC-Talk Adds Natural Language to StarCraft Multi-Agent Challenge for LLMs
- 66. Spectral transfer identity s=αγ ties curvature exponent to Hessian decay
- 67. ChatHealthAI Aligns Structured EHR Data with Frozen LLM for Clinical Reasoning
- 68. Study Explores Graph Scaffolds as Reasoning Aid for Large Language Models
- 69. NVIDIA releases Cosmos 3 with Super‑Text2Image and Nano‑Policy‑DROID
- 70. Guide: Run a Claude Managed Agent Task End‑to‑End via Session Stream
- 71. Microsoft unveils Surface NVIDIA RTX Spark Dev Box for AI agent development
- 72. Nvidia builds RTX Spark supercomputer chips with Microsoft for AI agents
- 73. LLM-derived valence direction aligns with EEG signals in 123 subjects
- 74. gSMILE Framework Tackles LLM Transparency by Mapping Prompt Responses
- 75. DAStatFormer extracts 24 ANOVA-selected features per channel, slashing data size
- 76. Test-Time Prompt Optimization Turns Demonstrations into Rewards for VLM Models
- 77. BitsMoE uses SVD to keep basis unquantized, allocating bits to expert spectral factors
- 78. Claude Code Leads Feature Set as Codex Adopts Similar Tools for Coding
- 79. MiniMax-M3 launches, beats GPT-5.5 and Gemini 3.1 Pro on benchmarks, costs 5‑10%
- 80. Turing Award winner Richard Sutton: Pure generative AI cannot do real science
- 81. Google I/O 2026 Showcases Gemini‑Powered Infinite Scaler and Code Countdown
- 82. Gemini App Targets General Users—Students, Writers, Marketers, and More
- 83. Study fine-tunes honest and deceptive variants of five transformers with LoRA
- 84. 3-large embedding wins 2.1 test; MiniLM wins 2.3; rerankers lag in 2.2
- 85. Proxy-Pointer RAG Bakes Emerson Deltas into Index for AT&T system
- 86. Study finds base AI models predict human behavior better than fine‑tuned chatbots
- 87. Chronos-2 uses known covariates such as weather for building demand forecasts
- 88. OpenAI upgrades GPT-5.5 readability, removes Canvas from Instant and Thinking
- 89. Deep learning models auto‑detect data features, reducing need for engineer input
- 90. Google's Gemini Spark sees my whole life, then friend‑zones my boyfriend
- 91. Researchers Find Failure Signatures in LLM Trading Agents' Planning Embeddings
- 92. SSD removes sync bottleneck in speculative decoding on MI300X
- 93. Claude Opus 4.8 Trained for Honesty, Flags Uncertainty, Reduces Frustrations
- 94. Transformer Architecture Reduces Perplexity by 2.92 vs Fine‑Tuning
- 95. Step 3.7 Flash runs on NVIDIA GPUs via SGLang, TensorRT-LLM, vLLM
- 96. LLMs Struggle with Causal Discovery While Interventional Agents Succeed
- 97. DynaSchedBench Introduces SESC and SSI to Rank LLM Scheduling Tasks
- 98. LLM-based Architecture Targets Explicit and Implicit Human Values in Text
- 99. Google AI launches Daily Brief in Gemini app for U.S. users 18+
- 100. Google Cloud unveils AI platform with Gemini, Wiz, Codemender to patch gaps fast